Sobre a etiologia da homossexualidade

The most publicized research studies on the causes of homosexuality have been those examining the role of genetics. Indeed, a number of studies have attempted to assess the heritability of the homosexual orientation. The primary way of examining the genetic contribution to a given behavior or disposition such as sexual orientation is through twin studies. The premise is that by comparing monozygotic (MZ) twins (who share 100% of their genetic code), dizygotic (DZ) twins and biological siblings (who both share approximately 50% of their genetic code), and adopted siblings (who have none of their genetic code in common), researchers can glean an understanding of whether a trait is heritable and to what extent (Billings & Beckwith, 1993).

However, in order to truly assess the effects of genes, one has to control or neutralize the contribution of environmental factors in the development of the characteristic in question. Such environmental factors could include anything from position in the womb, maternal nutrition, and the hormonal milieu of the uterine environment to post-natal care, early illnesses, parenting style, family constellation, education, socialization, and so forth. These influences can have profound effects not only on a person’s behavior but even on their biological make-up.

Thus, researchers are coming to recognize that the hard division of nature—nurture is no longer tenable. It is always a “both—and”, especially in traits as complicated as sexual orientation. It is important to note that several of the factors listed above pertain to the prenatal environment. In this period of development, the environment can have profound effects on the organism, and this will be discussed further below.

In the most desirable type twin study, researchers examine MZ twins who were separated at birth and reared apart. This is considered the ideal in behavioral genetics. As the reasoning goes, since such twins have the same genetic code but are reared in different environments, any behavioral similarities they manifest likely have a strong genetic basis. Unfortunately, these situations are rare, and they are extremely so if the trait being studied is itself uncommon. Homosexuality is just such a case.

A great deal of research in recent years has shown that roughly 2-3% of men in the United States are homosexual (Fay, Turner, Klassen, & Gagnon, 1989; Rogers & Turner, 1991; Leigh, 1993; Billy, Tanfer, Grady, & Klepinger, 1993; Seidman & Reider, 1994; Black, Gates, Sanders, & Taylor, 2000). The figure is even smaller for women, approximately 1-2% (Diamond, 1993; Laumann, Michael, Gagnon, & Michaels, 1994; Wellings, Field, Johnson, & Wadworth, 1994).

Thus, locating MZ twins who were separated at birth and of whom at least one twin developed a homosexual orientation is unfeasible. Moreover, such a twin study (MZ twins reared separately) does not in fact neutralize the environment to the extent that researchers surmise. The reason for this is that MZ twins, even if they are separated from the moment of birth forward, still shared the same, profoundly influential environment for the previous nine months. Thus, every twin study, no matter how elegant in design, cannot fully tease apart the effects of genes and environment.

Researchers try to cope with these difficulties by comparing MZ twins who were reared together with other sibling pairs such as DZ twins, normal biological siblings, and adopted siblings with no biological relation between them. Bailey and Pillard (1991) followed this general paradigm by examining family patterns of adult males with homosexual orientation who had either a MZ twin, DZ twin, or adopted brother. These researchers found a concordance rate (if one twin was homosexual, the other was as well) of 52% among the male MZ twins who were reared together in the study. The figure for female MZ twins was 48%.

Likewise, male DZ twins reared together showed a concordance rate of 22% (16% for women) but this was not significantly different from the rate for adoptive brothers (Bailey & Pillard, 1991). What this study shows is that there may be a genetic component to homosexuality, but there appear to be substantial environmental factors as well. Why else would the concordance rate among MZ twins be so much less than 100% and why would there be any concordance at all among adoptive brothers?

Though far from conclusive, this study by Bailey and Pillard was highly influential. Several such twin studies followed, and, taken together, the lowest concordances found for homosexuality among MZ twins was 47% for men and 48% for women (Bailey, Dunne, & Martin, 2000). However, the samples of twins included in these studies were largely recruited through advertising in gay or lesbian publications. This creates the possibility of “ascertainment bias”.

In other words, “twins deciding whether to participate in a study clearly related to homosexuality probably considered the sexual orientation of their co-twins before agreeing to participate” (Bailey et al., 2000, p. 533). This non-random sampling, of course, would result in biased data. Bailey, Dunne, and Martin (2000) overcame this methodological weakness by using a large sample (N=4901) of adult twins recruited from the Australian Twin Registry. In that study, the researchers identified 27 pairs of male twins in which at least one of the men was homosexual. However, only 3 of these pairs were concordant (both twins were homosexual).

Thus, 3 of 27 pairs of male twins (approximately 11%) were concordant on homosexuality in this systematic study. Likewise, only 3 of 22 pairs of female twins (approximately 14%) were concordant. These researchers conclude, “These rates are significantly lower than the respective rates for the two largest prior twin studies of sexual orientation…. This suggests that concordances from prior studies were inflated because of concordance-dependent ascertainment bias” (Bailey, et al., 2000, p. 533). They go on to state,

Consistent with several studies of siblings… we found that sexual orientation is familial. In contrast to most prior twin studies of sexual orientation, however, ours did not provide statistically significant support for the importance of genetic factors for that trait (p. 534).

This does not rule out the possibility that there is a genetic component to homosexuality but rather suggests that “sexual orientation is inherited, if at all, in a complex manner” (Bailey & Pillard, 1995, p.144, emphasis added).

A second line of research examining the supposed genetic basis for homosexuality employs the molecular analysis of blood samples from relatives of homosexuals and a method called linkage analysis, which maps genes onto chromosomal regions, to identify the specific gene or genes that influence sexual orientation. Based on evidence that “gay men have more homosexual uncles and cousins through the maternal line than on the paternal side” some researchers have attempted to identify an X-linked gene that influences male sexual orientation (Rahman & Wilson, 2003, p. 1342).

Hamer, Hu, Magnuson, Hu, and Pattatucci, (1993) initially reported findings consistent with this hypothesis that implicated the chromosomal region Xq28 in the heritability of male homosexuality. A similar though weaker relationship was reported later by members of the same research group (Hu et al., 1995).

However, the methods used in these studies and therefore the reported results have been called into question. Risch, Whleeler, and Keats (1993) criticized the research design and statistical methods used by Hamer et al. (1993) and asserted that their conclusions might not be supported by the data. In addition, Rice et al. (1999) attempted to replicate the previously reported link between Xq28 and male homosexuality using similar methods in independently conducted genetic studies, but their results did not support the Xq28 hypothesis. Indeed, these researchers stated,

It is unclear why our results are so discrepant from Hamer’s original study. Because our study was larger than that of Hamer et al., we certainly had adequate power to detect a genetic effect as large as was reported in that study. Nonetheless, our data do not support the presence of a gene of large effect influencing sexual orientation at position Xq28 (Rice et al., 1999, p. 668).

These issues were again revisited in the journal Science in 1999 (Hamer et al., 1999; Rice, Risch, & Ebers, 1999). It is clear from these articles that because scientists are treading on new methodological ground, there is no simple solution to this question. The methods being used in these studies have not been tested and verified in studying a trait as complex as sexual orientation. Hamer et al. (1999) defended their approach and attempted to meta-analyze four studies in this area concluding that they collectively support the Xq28 hypothesis.

However, Rice, Risch, & Ebers (1999) pointed out that if only those studies carried out by independent investigators are considered, which is necessary to reduce potential biases, no researchers outside Hamer’s own group has found support for the Xq28 hypothesis: “Thus, the conclusion remains that the original studies of Hamer and colleagues are not replicated” (Rice, Risch, & Ebers, 1999, p. 806).

A study conducted by Bailey et al. (1999) also examined the hypothesis that homosexuality is the result of an X-linked gene using more stringent standards in recruitment analysis than many previous studies. Three sampling techniques were utilized: recruitment from an HIV clinic, a gay pride parade, and through homophile publications. All participants were interviewed about familial patterns of non-heterosexuality. Only the sample attained through magazine advertisement knew of this purpose of the study. The researchers found a rate of occurrence of homosexuality among brothers of a homosexual male ranging from 7.3% to 9.7%.

This suggests a modest familial (not necessarily genetic) component to the origin of male homosexuality. Bailey et al. (1999) also found a slight increase in the appearance of female homosexuality among sisters of gay men. However, the familial patterns of homosexuality observed in this study did not support the Xq28 hypothesis. “This study found no evidence that male sexual orientation is influenced by an X-linked gene” (Bailey et al., 1999, p. 84).

One final study worth mentioning was recently conducted by Mustanski, DuPree, Nievergelt, Bocklandt, Schork, and Hamer (2005). Using blood samples of 456 individuals from 146 unrelated families, the researchers conducted for the first time a search of the entire human genome for a genetic basis to male homosexuality. Though three gene regions resulted in elevated values, none were large enough to generate a statistically significant result. In addition, they found no evidence of a link to the Xq28 region. Thus, as stated above, research into the genetic basis for homosexuality, taken as a whole, is inconclusive. While there may be a modest heritability to such an orientation, the existence of an overriding gay gene seems highly unlikely at this time.

Another biologically based line of research bearing on the question of the etiology of homosexuality deals with the pre-natal environment. As already alluded to, fetal development is a remarkably crucial time in the development of the human organism. In particular, brain systems develop rapidly during this period and are quite sensitive to hormonal and other biochemical influences.

Normally in a male child, his testes will begin producing testosterone during the fetal period which masculinizes various body structures and systems. (Girls produce androgens as well from their adrenal glands, though usually in much smaller quantities than boys who produce them in their testes and adrenals.) Besides the obvious differences such as genitalia, musculature, etc., this masculinzation process affects the development of various brain structures and, thus, their functions (see chap. 8 of Baron-Cohen, 2003). The development of the child in the womb can also be influenced by the mother’s hormone levels. It has been hypothesized that abnormal hormone levels or other deviations during the fetal period could be responsible for the development of non-heterosexual orientations (for reviews, see Mustanski, Chivers, & Bailey, 2002; Rahman & Wilson, 2003).

Researchers have made various attempts to study such an effect upon the development of homosexual orientations. While for ethical reasons, experimenters cannot manipulate the hormonal milieu of the prenatal environment in humans, they can study cases in which such abnormalities occur naturally to see if a homosexual orientation is more likely to result. In addition, a certain amount can be learned from animal studies. Finally, somatic and neuropsychological variables known to be strongly affected by prenatal hormones can be compared in adults to see if they differ according to sexual orientation.

All of these methods have been utilized, and it appears that in some respects, female homosexuals appear to have experienced increased prenatal androgenization compared to heterosexuals (more masculine auditory systems and waist-to-hip ratio, higher salivary testosterone levels, less desire to give birth, etc.). Likewise, on some measures homosexual men appear to have been less androgenized than heterosexual men (more female-like pubertal onset, weight, height, cognitive spatial abilities, and language abilities; higher rates of non-right handedness).

However, on other measures homosexual men appear hypermasculine, i.e. exposed to greater quantities of androgens (more masculine relative finger lengths, possibly larger penises [according to two studies: Nedoma & Freund, 1961; Bogaert & Hershberger, 1999], and possibly hypermasculine auditory systems [see McFadden, 2002]). Some differences in the size and shape of certain adult brain structures have also been identified across sexual orientation categories, but it is unknown whether these neural differences cause or are consequent upon homosexual lifestyles. After all the brain is a “plastic” structure that changes throughout life in response to one’s genetic/biological endowment, experiences, and behaviors.

Thus, the data on the role of prenatal androgens are very complex, and no scholarly consensus exists on how to integrate and interpret them. If abnormal androgen levels are what drive these differences, they appear to affect males and females differently since homosexual men and homosexual women do not always differ from their heterosexual counterparts on the same variables.

Moreover, the data on homosexual men are mixed. Rahman and Wilson (2003) put forward the hypothesis that perhaps a genetic factor in some men alters the distribution of androgen receptors in various brain structures. Thus, when exposed to prenatal androgens at whatever level, the result will be that some brain structures will be more masculinized than normal and others will be less masculinzed than normal because of the redistribution of the receptors. While this makes a good deal of sense, the research has not yet shown this to be the case.

Perhaps the most well replicated phenomenon related to the causes of homosexuality is the “fraternal birth order effect”. To put it simply, “Several studies, which collectively examined over 7000 subjects, have shown that homosexual men have, on average, a greater number of older brothers than do comparable heterosexuals” (Ellis and Blanchard, 2001, p. 543).

Thus, something about being born later in a line of several brothers seems to increase the likelihood of male homosexuality. Researchers believe that the fraternal birth order effect per se is not the ultimate causal factor of male homosexuality in these cases but that there is some other mechanism in place that mediates the effect. However, what this causal mechanism is is still highly debated. Most favor a biological explanation such as a fluctuation in prenatal androgens or increased maternal immune system response related to multiple, successive male pregnancies.

Another possibility with mixed empirical support suggests that maternal stress, which could increase with the births of multiple sons, may alter fetal development in such a way as to increase the likelihood of a homosexual orientation. Stress hormones are in fact produced in the adrenal glands (and are therefore, androgens or at least androgen-like) and can interfere with normal fetal development.

Other explanations for the fraternal birth order effect are psychosocial in nature and posit such potential causes as ostracization of boys by older brothers or increased early sex play among boys with several older brothers (see James, 2004a; 2004b). Presently, however, the fraternal birth order effect remains largely unexplained in the absence of any overwhelmingly conclusive evidence.

Moreover, it is important to note that according to researchers only 14.8% to 15.2% of homosexual men can attribute their orientation to this effect (Cantor, Blanchard, Paterson, & Bogaert, 2002). Though it is a well established and easily identifiable phenomenon connected with the genesis of male homosexuality, the fraternal birth order effect does not appear to be the primary cause of homosexuality in the majority of gay men. Nothing analogous to this effect has been found in women.

In light of the foregoing, it is important to recall that sexual orientation does not emerge immediately upon conception and birth but takes time to develop throughout childhood, adolescence, and even adulthood. Thus, there is a great deal of room for experience to affect this process. Consequently, researchers investigating the various factors discussed above agree that rather than causing homosexual orientation directly these influences likely precipitate a set of pre-homosexual traits and dispositions that in turn increases the likelihood of adult homosexuality (Bailey & Zucker, 1995; Dunne, Bailey, Kirk, & Martin, 2000).

The term coined for this is childhood gender nonconformity (CGN), which simply refers to a certain conglomeration of sex atypical traits (physical, psychological, and behavioral features). CGN is known to be consistently associated with later homosexual orientations as well as many of the biological and familial factors discussed above (Bailey & Zucker, 1995). However, since not everyone with CGN develops an adult homosexual orientation, there must be some environmental influences that contribute to such a development. These could include for instance failed relationships with one’s parents and/or peers.

Indeed, classic psychodynamic theory held that male homosexuality was primarily caused by a maladaptive family pattern in which the father is weak and distant and the mother is over-controlling (roughly the reverse is posited for females). While there has been very little data in the last 20 years supporting these specific claims, research does emphasize the importance of relationships in the development of sexual orientation. For instance, an interesting study by Landolt, Bartholomew, Saffrey, Oram, and Perlman (2004) examined the relationship between CGN and rejection from parents and peers in a sample of 191 gay and bisexual men. They found, “Gender nonconforming behavior in childhood was associated with maternal, paternal, and peer rejection” among these gay and bisexual men (p. 124).

Another recent study, though methodologically weak, found that among a sample of Catholic seminarians, those admitting a homosexual orientation also reported “more emotional distance from their fathers than heterosexual seminarians” (Seutter & Rovers, 2004, p. 46).

Bem (1996; 2000) proposed a model of sexual orientation development which he dubbed the “exotic becomes erotic” theory. He proposes that biological variables influence childhood temperaments which in turn affect a child’s degree of gender conformity. A child who has nonconforming traits and behaviors feels him/herself to be different from peers. This “feeling different” from same-sex peers can evolve into an erotic attraction in adolescence:

An individual’s protracted and sustained experience of feeling different from same- or opposite-sex peers throughout childhood and adolescence produces a correspondingly sustained physiological arousal that gets eroticized when the maturational, cognitive, and situational factors coalesce to provide the critical defining moment (Bem, 2000, p. 539).

While these various accounts all have merit to them and a certain ring of truth, one aspect of the development of a homosexual orientation not explicitly tapped in any of them is the role of child sexual abuse (CSA) and other early sexual experiences. There seems to be a link between CSA perpetrated by men (or older males) against boys that predisposes these boys to later identify as homosexual.

Indeed, one reviewer states, “The effects of sexual victimization on male children, placed in the context of the prevailing concept of masculinity, include mental disorders, the probability of becoming rapists and incest offenders as adults, and the development of homosexual identification” (Vander Mey, 1988, p. 61). Several studies support this claim. For instance, Finkelhor (1984) found a statistically significant relationship between CSA victimization and later homosexual activity in adulthood (cited in Dimock, 1988). In a study of 25 adult males who had been sexually abused in childhood, the majority of these men experienced “masculine identity confusion” characterized by “confusion regarding sexual preference and… male roles” (Dimock, 1988, p. 208). Homosexual feelings and/or behaviors were common among these men. Noting that “the literature reports that there may be a significant number of gay men who have been sexually abused as children,” Schwartz (1994) presented clinical data from eight men who had been victims of CSA. “Six of the eight men identified themselves as homosexual in their sexual identity; two were heterosexual. Of the two heterosexuals, one continuously questioned his sexual orientation” (Schwartz, 1994, p. 182).

Likewise, Johnson & Shrier (1985) compared 40 male victims of CSA and an age-matched control group and found that 47.5% of the CSA victims went on to develop a homosexual orientation and another 10% became bisexual. “The study group identified themselves as currently homosexual nearly seven times as often and bisexual nearly six times as often as the control group” (Johnson & Shrier, 1985, p. 374).

These small studies finding an increased likelihood of homosexual identification in male victims of CSA are corroborated by at least twelve more recent and systematic ones that have all documented elevated rates of CSA among homosexual men (Saewyc, Pettingell, & Skay, 2004; Kalichman et al., 2004; Ratner et al., 2003; Garcia, Adams, Friedman, & East, 2002; Dolezal & Carballo-Diéguez, 2002; Tomeo et al., 2001; Paul et al., 2001; Krahé et al., 2001; Paris et al., 1995; Doll et al., 1992; Baier et al., 1991; Haverkos, Bukoski, & Amsael, 1989; for reviews see Relf, 2001; Holmes & Slap, 1998). For instance, using a nonclinical sample of 942 participants, Tomeo et al. (2001) found that 46% of homosexual men had been sexually abused as a child by a man. This was significantly greater than the 7% rate found among heterosexual men in that study. Among a sample of 307 Latin American men who have sex with men (MSM), Dolezal and Carballo-Diéguez (2002) observed “that early sexual contact [with males] is common among these men.

Fifty-nine percent had had some sexual/genital contact prior to their 13^th birthday. In the majority of those cases, they had a partner who was at least 4 years older than they were” (p. 169-170). Finally, in a systematic study of 2881 MSM, Paul et al. (2001) found that one fifth had experienced CSA. They state,

Our study confirms and extends prior research indicating high prevalence levels of childhood sexual abuse (CSA) among MSM. Such prevalence levels might be higher if we had elicited data about experiences involving non-contact sexual victimization (i.e., sexual exposure or exhibitionism). Overall, these men’s CSA experiences were characterized by high levels of penetrative sex, physical force, and perceptions of these events as distressing (Paul et al., 2001, p. 575).

Interestingly, some research suggests that boys who have gender nonconforming traits such as effeminate physical features are rated as more attractive by adults than a control group of typical boys (Zucker, Wild, Bradley, & Lowry, 1993).

Even in cases where such young men and boys do not report the experience as abusive, early sexual contact with other males is common prior to identifying oneself as gay. For instance, Dawood et al. (2000) found that one third of their sample of gay men with gay brothers had engaged in sexual activity with their siblings in childhood. “Among the 21 participants who indicated that some form of sex play occurred, levels of activity included touching and mutual masturbation (N=16), giving or receiving fellatio (N=9), and anal intercourse (N=4)” (p. 161). A recent study of 961 Dutch gay and bisexual men “found that 68% of respondents engaged in their first same-sex experience before coming-out” (Schindhelm & Hospers, 2004, p. 585). Another study found that 58% of homosexual and bisexual men came to their sexual identity through a sex-centered sequence, i.e. one in which same-sex activity preceded their identification as gay or bisexual (Dubé, 2000).

In a study comparing homosexual men and women, researchers discovered that homosexual men were more likely to pursue sex before identifying themselves as gay, whereas for women the context of their homosexual identity development was more emotionally oriented (Savin-Williams & Diamond, 2000).

This last finding raises an interesting but challenging issue: sexual orientation in men and women is markedly different and develops in divergent ways. Summarizing research in this area, Diamond (2003) states,

Women show greater variability than men in the age at which they consciously question their sexuality, and the age at which they pursue their first same-gender sexual contact…. Also, women place less emphasis on the sexual component of their lesbian or bisexual identification, both during and after the questioning process…, and are more likely to report that their sexuality is fluid and chosen versus fixed and biologically given… (p. 185).

Thus, for men, a homosexual orientation seems to involve a strong inclination to engage in sexual activity with other men. Whereas, for women the question of sexual orientation is much more caught up in affectional bonding and is more often a matter of conscious choice than an irresistible urge.

This review of the research shows us that many known factors (and there are likely still more unknown factors) are implicated in the development of homosexuality, but no one factor seems powerful enough to trump them all. Rather, the adult homosexual has probably arrived at that orientation through a complex, idiosyncratic combination of biological, experiential, and volitional factors. This process seems clearly divergent for men and women and likely differs from individual to individual among gays and among lesbians.

Andrew J. Sodergren, Psy.D.